A Perspective on Models and the Quality Sciences: Some Challenges and Future Directions – Winter 1999 Youden Address

Abstract: A walk through the history of the evolution of theories about our physical world will reveal the nature of scientific thought. From Aristotle’s postulate around 340 B.C. that the Earth was a sphere to Richard Feynman’s postulation of quantum electrodynamics in the 20th century, each scientist took the model or theory that had been proposed and improved or replaced it based on empirical observations. Statistical process theory is no different. The Shewhart model used in Shewhart control charts seeks to monitor deviations from an output in a given (industrial) process. As such, the Shewhart model should be subjected to the same criticism, refinement and even reformulation given to models and techniques in other fields. The best way to evaluate a model is to subject the hypothesis to rigorous experimentation, repeated measurements and peer-review.

Keywords: Control charts - Statistical process control - Time Series Data - Statistical Methods - Industrial experiments - Inductive learning